From owner-freebsd-questions Sat Jul 8 12:24:53 2000 Delivered-To: freebsd-questions@freebsd.org Received: from hawk.prod.itd.earthlink.net (hawk.prod.itd.earthlink.net [207.217.120.22]) by hub.freebsd.org (Postfix) with ESMTP id D5E9B37B896 for ; Sat, 8 Jul 2000 12:24:49 -0700 (PDT) (envelope-from egravel@earthlink.net) Received: from earthlink.net (hybrid-024-221-117-152.phoenix.speedchoice.com [24.221.117.152]) by hawk.prod.itd.earthlink.net (8.9.3-EL_1_3/8.9.3) with ESMTP id MAA03121 for ; Sat, 8 Jul 2000 12:24:48 -0700 (PDT) Message-ID: <39678098.698CEF07@earthlink.net> Date: Sat, 08 Jul 2000 12:27:20 -0700 From: Emmanuel Gravel X-Mailer: Mozilla 4.61 [en] (X11; I; FreeBSD 4.0-RELEASE i386) X-Accept-Language: en MIME-Version: 1.0 To: freebsd-questions@FreeBSD.ORG Subject: Re: Performance issues with dd References: <200007081838.LAA27699@implode.root.com> Content-Type: text/plain; charset=us-ascii Content-Transfer-Encoding: 7bit Sender: owner-freebsd-questions@FreeBSD.ORG Precedence: bulk X-Loop: FreeBSD.ORG I was trying to avoid RAID if I could. The reason being, if the HD fails, I don't want any CPU overhead (even though I've got some to spare) and getting back up would only require powering down and removing a jumper. I don't think there'd be any filesystem changes at the time I'm doing the dd. It's on a closed network, and no one's around when I'm doing it. I've used tar but I don't see how it would let me create a complete HD image on the other drive. I see it being useful for creating backup tapes, or backup archives, but a full image? I guess I'll have to look deeper in the man pages... I'll look into block size too. Any way to force the HD's to do write caching? I'm pretty sure it's implemented in hardware, but is there anything that would disable it? Thanks for all the input! Emmanuel David Greenman wrote: > > > > As a backup measure, I'm using two hard drives of the same size, > > > but only mounting one, and I have a cron job that dd's one to > > > the other every night. The drives are 18 GB U160 SCSI, on a > > > AHA-29160N. The whole process takes 45 minutes. A quick calc > > > tells me it should be finished at best in 2 minutes, at peak > > > rate. This is over a 20x difference. > > > > > > Does anyone have any idea why it's taking so long? Is it just > > > dd that's the problem? The CPU is a PIII 650, and there's more > > > than enough memory to go, so I don't think the CPU would be > > > any kind of an impact here. Is there a better app out there > > > to do the job? I'm dumping one full raw device to the other, > > > not individual partitions. Could that be another problem? > > > >Since you're duplicating an entire disk why not try RAID 0? It's > >probably a better solution than "dd" and will keep you spare as current > >as possible. > > You mean RAID 1 (mirroring). RAID 0 is just striping and offers no > redundancy. > > >As far as "dd" goes, consider ditching it and using dump/restore, pax, > >or tar. You may not know it, but "dd" is duplicating the entire disk, > >bad sectors and all. > > Worse than that, if the filesystem changes during the dd, you could end up > with a copy that is completely trashed and worthless. It's really much smarter > to do as you suggest and use something else to copy the contents. I personally > prefer tar. To Unsubscribe: send mail to majordomo@FreeBSD.org with "unsubscribe freebsd-questions" in the body of the message